noahpunintended/picard:1.7b-q8

noahpunintended/ picard:1.7b-q8_0

452 Downloads Updated 1 year ago

smaller llms fine tuned to be a voice assistant

ollama run noahpunintended/picard:1.7b-q8_0

curl http://localhost:11434/api/chat \
  -d '{
    "model": "noahpunintended/picard:1.7b-q8_0",
    "messages": [{"role": "user", "content": "Hello!"}]
  }'

from ollama import chat

response = chat(
    model='noahpunintended/picard:1.7b-q8_0',
    messages=[{'role': 'user', 'content': 'Hello!'}],
)
print(response.message.content)

import ollama from 'ollama'

const response = await ollama.chat({
  model: 'noahpunintended/picard:1.7b-q8_0',
  messages: [{role: 'user', content: 'Hello!'}],
})
console.log(response.message.content)

Details

Updated 1 year ago

1 year ago

8a2e9d7fb8bd · 1.8GB ·

model

archllama

parameters1.71B

quantizationQ8_0

1.8GB

params

{ "stop": [ "<|im_start|>", "<|im_end|>" ], "temperature": 0 }

75B

template

{{ if .System }}<|im_start|>system {{ .System }}<|im_end|>{{ end }} <|im_start|>user {{ .Prompt }}<|

131B

Readme

Please use noahpunintended/picard:0.36b-f16 specifically for best performance :-).

Instructions how I did this: https://github.com/nkasmanoff/ft-flow

Used in this repo: https://github.com/nkasmanoff/pi-card

System prompt:

“You are Pi-Card, the Raspberry Pi AI assistant.”

Again, please use 0.36b-f16for best results.